Machine vision or inspection systems have become a vital component in integrated manufacturing systems. They can sort, package, and perform defect analysis without human intervention. For instance, by inspecting holes being drilled the system can determine if a drill bit is worn.
Most machine vision systems are based upon digital electronic technology that uses serial or one dimensional processing. An image is captured and stored as a matrix of electrical signals. The image is then preprocessed to enhance edges, improve contrast, and otherwise isolate the object to be recognized. A comparison function compares the enhanced image to one or more stored reference images. These preprocessing and comparison functions are typically performed by standard microelectronic, digital equipment on a bit-by-bit or vector basis. Accordingly, the techniques are typically serial and inherently one dimensional, whereas the images being processed are two dimensional. This dichotomy results in very intensive processing requirements, is particularly difficult for one dimensional digital equipment, and, even with an extraordinary amount of memory capacity and processing capability takes a relatively long time to complete. Digital processing hardware has been enhanced and the software and algorithms have been improved over prior art machine vision systems. However, these improvements have come at the expense of additional system complexity, system costs and programming complexity, and still suffer from the inherent limitations of serial processing.
In some systems, the image to be processed is converted into a Fourier or other transform. The Fourier transform presents information about the image of the object in a very useful, symmetrical pattern which represents the object in terms of its spatial frequencies. However, the calculation of a Fourier transform on a digital computer is extremely intense, and may take a computer as powerful as a Micro Vax II about a minute to complete. Even powerful and expensive state of the art array processors take a full second to merely produce the transform. In modern industrial plants, the production line rates are often a full order of magnitude faster than this.
The computational intensity and time are significantly reduced using parallel processing techniques, such as those available when the real image of the object undergoing inspection is converted to a transform image and optically processed. Following generation of the transform image it is "processed" by quantifying the light from a preselected number of spatial domains or segments of the transform image. These quantities are then electronically processed to provide an abbreviated or composite characteristic signature of the transform image, and thus of the object upon which it is based. In comparison to the time and expenses involved when dealing with entire transform images, the signatures may be rapidly and economically obtained and then evaluated to determine whether the object does or does not conform to preselected standards.
Although possessing the above-noted benefits, an inspection system of the foregoing type must include signature-generating means, in addition to the other system components such as means for generating electrical signal data representative of the appearance of each inspected object, means for receiving such signal data and producing a visual image of each object represented thereby, and means for producing a transform image of the object from the aforesaid visual image. The successful utilization of such an inspection system in association with a high-speed production line or the like, wherein the objects must be inspected quite rapidly, requires rapid and accurate control and synchronization of the operation of the various system components.